Fractal stochastic processes provide a rich class of models that are useful in understanding biomedical time series and images. In the past year, we have investigated a Gaussian spectral synthesis method for simulating fractal time series and used this method to demonstrate the nonconvergence of certain correlation dimension estimators for power law processes (convergent results had been reported using a flawed spectral synthesis method).